Oculoplastic surgeons face a dilemma whether or not to discontinue anticoagulants before ptosis surgery. Based on consultation with a cardiologist, most oculoplastic clinicians recommend discontinuing anticoagulants prior to ptosis surgery, but the effect on postoperative outcomes is still unknown.
Previous studies investigated the risk of anticoagulant treatment on ophthalmic surgery, but a majority of them analyzed cataract or retinal surgery, with limited focus on oculoplastic surgery. Furthermore, the effect of anticoagulants on oculoplastic procedures is still controversial, and there is no valid guideline.
Recent studies [5–8] revealed that anticoagulant treatment has no significant effect in most oculoplastic surgery. Bartley  reported no major bleeding complication in patients treated with anticoagulants who underwent oculoplastic procedures, including blepharoplasty and dacryocystorhinostomy (DCR). The American College of Chest Physicians (ACCP) classified eye surgery including oculoplastic procedure as ‘low risk’, which is defined by a probability of 0 to 2% of major bleeding events occurring within 2 days of surgery and recommended continuing treatment with anticoagulants perioperatively. Similarly, our previous study  revealed that the use of intraoperative ketorolac, known as blood thinner, in oculoplastic surgery does not increase postoperative bleeding risk. Other studies also revealed the low incidence of severe complications in patients undergoing anticoagulant therapy and non-significant differences compared with control group. Philip et al  analyzed the intraoperative and postoperative complications in patients undergoing oculoplastic procedure, and reported that the proportion of major complications affecting the surgical outcome was only 0.4% and found no statistically significant difference among patients with and without taking anticoagulants. As new anticoagulants, including direct oral anticoagulant (DOAC), previously known as new oral anticoagulants (NOAC), are routinely recommended, several studies have investigated the half-lives of DOACs. A study published in the European Heart Journal showed that DOACs have comparable half-lives and suggested that their residual anticoagulant effects do not differ or persist upon treatment interruption.
However, our results were inconsistent with their results, and confirmed the effect of anticoagulants. Interestingly, our study showed significant ecchymosis and edema after surgery despite adequate discontinuation of anticoagulants. The degree of postoperative ecchymosis was severe in the group that discontinued with anticoagulants before surgery at postoperative one week and one month despite proper discontinuation. In addition, postoperative MRD1 was significantly lower in group 1 at one week.
Similarly, traditionally even if anticoagulants are discontinued appropriately prior to oculoplastic surgery, the risk of severe complications such as retrobulbar hemorrhage may still remain. In addition, according to questionnaire surveys conducted by Esparaz et al , the majority of oculoplastic surgeons experienced various intraoperative or postoperative complications in patients with a history of anticoagulant usage. Parkin et al argued the risks of taking anticoagulants prior to blepharoplasty, indicating that the incidence of ocular hemorrhage was high in patients who took anticoagulants.
While these contrasting studies have focused on intraoperative and postoperative complications in patients taking anticoagulants, the effects of such complications on surgical outcomes were unclear. Furthermore, especially in patients undergoing blepharoplasty, it is necessary to focus on more common complications such as postoperative ecchymosis, edema or under-correction rather than major complications such as retrobulbar hemorrhage. Therefore, the aim of our study was to determine the effect of discontinue anticoagulant use before surgery on the degree of ecchymosis after surgery and the effect on the surgical outcome.
The possible mechanisms supporting our findings are as follows. First, irreversible hematologic or microvascular changes may have occurred due to hypertension and long-term use of anticoagulants. Gregory et al  showed that hypertension was associated with abnormal hemostasis/fibrinolytic function. Wall et al  reported that hypertension was a predictor of permanent vascular change. In our study, the prevalence of hypertension was significantly higher in the group exposed to anticoagulants, which may have altered the postoperative outcome. Second, the anticoagulant action would have persisted even when the anticoagulant was discontinued. Two studies [16, 17] reported residual serum concentrations of anticoagulants in patients who underwent elective invasive procedures. Anne et al  confirmed persistent anticoagulant effects in most patients even after discontinuing for 3 times the half-life of the known anticoagulant. Another possible hypothesis is that delayed bleeding occurred due to re-administration of anticoagulant after surgery. In fact, the challenge of delayed bleeding after endoscopy-like procedures in patients taking anticoagulants is being debated worldwide. Hideomi et al.  argued the possible risk of delayed bleeding in patients taking anticoagulants. Harada et al.  also reported a 9% higher rate of delayed bleeding. Similarly, our study confirmed that the clinical application of the currently known half-lives of anticoagulants in periprocedural management is unstable. Therefore, in line with the emergence of new anticoagulants and their increased use, further active studies investigating periprocedural management are needed.
Unfortunately, our study has some limitations that have yet to be clearly addressed. First, this study was retrospective with a relatively small sample size, which may limit the statistical power. Second, no objective indicators were available to quantitatively evaluate the degree of bleeding experienced by patients. Lastly, the ecchymosis grade and persistence can be proxy and subjective despite evaluation by 4 masked graders to minimize bias.
Otherwise, it is important to balance the risk of potential intraoperative bleeding and postoperative ecchymosis against the risk of systemic thromboembolic events. Therefore, the oculoplastic surgeon must establish proper guidelines for patients taking and discontinuing anticoagulants by evaluating the possible complications and surgical outcomes. Also, the possible complications after surgery need to be borne in mind, even if the drugs are discontinued properly. Fortunately, the patients should be reassured and monitored during the intervention and subsequently. Even if such ecchymoses and edema occur, there may be no significant effects on the surgical outcome.
In conclusion, postoperative ecchymoses were more severe in group 1 by one month after ptosis surgery even though after discontinuing with anticoagulants. Surgeons should be careful about this before operation.